Import libaries

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns 
from xgboost import XGBClassifier # XGboost is used to achieve high accuracy on predictions.

Read Dataset

df=pd.read_csv('NIFTY 50_Data.csv')
df

	Date	Open	High	Low	Close
0	02-Mar-22	16593.10	16678.50	16478.65	16605.95
1	03-Mar-22	16723.20	16768.95	16442.95	16498.05
2	04-Mar-22	16339.45	16456.00	16133.80	16245.35
3	07-Mar-22	15867.95	15944.60	15711.45	15863.15
4	08-Mar-22	15747.75	16028.75	15671.45	16013.45
...	...	...	...	...	...
288	03-May-23	18113.80	18116.35	18042.40	18089.85
289	04-May-23	18081.00	18267.45	18066.70	18255.80
290	05-May-23	18117.30	18216.95	18055.45	18069.00
291	08-May-23	18120.60	18286.95	18100.30	18264.40
292	09-May-23	18303.40	18344.20	18229.65	18265.95

293 rows × 5 columns

df.head()

	Date	Open	High	Low	Close
0	02-Mar-22	16593.10	16678.50	16478.65	16605.95
1	03-Mar-22	16723.20	16768.95	16442.95	16498.05
2	04-Mar-22	16339.45	16456.00	16133.80	16245.35
3	07-Mar-22	15867.95	15944.60	15711.45	15863.15
4	08-Mar-22	15747.75	16028.75	15671.45	16013.45

EDA

'Information of dataset using info function'
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 293 entries, 0 to 292
Data columns (total 5 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   Date    293 non-null    object 
 1   Open    293 non-null    float64
 2   High    293 non-null    float64
 3   Low     293 non-null    float64
 4   Close   293 non-null    float64
dtypes: float64(4), object(1)
memory usage: 11.6+ KB

'Central tendency of dataset using describe function'
df.describe()

	Open	High	Low	Close
count	293.000000	293.000000	293.000000	293.000000
mean	17348.698976	17432.734642	17241.851706	17341.216041
std	800.296618	785.375343	808.989696	799.308928
min	15272.650000	15382.500000	15183.400000	15293.500000
25%	16900.650000	17037.600000	16788.600000	16887.350000
50%	17519.400000	17625.550000	17429.700000	17530.850000
75%	17920.850000	18022.800000	17818.400000	17915.050000
max	18871.950000	18887.600000	18778.200000	18812.500000

'Check null values in dataset'
df.isnull().sum()

Date     0
Open     0
High     0
Low      0
Close    0
dtype: int64

'Convert datatype of date to datetime'
df['Date']=pd.to_datetime(df['Date'])

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 293 entries, 0 to 292
Data columns (total 5 columns):
 #   Column  Non-Null Count  Dtype         
---  ------  --------------  -----         
 0   Date    293 non-null    datetime64[ns]
 1   Open    293 non-null    float64       
 2   High    293 non-null    float64       
 3   Low     293 non-null    float64       
 4   Close   293 non-null    float64       
dtypes: datetime64[ns](1), float64(4)
memory usage: 11.6 KB

'Calculate the daily range (high - low)'

df['Range'] = df['High'] - df['Low']

'Calculate the average range'

avg_range = df['Range'].mean()
print('Average daily range:', avg_range)

Average daily range: 190.88293515358356

Graphical Representation

'Plot open rate as with respect to date'
plt.figure(figsize=(10,5))
plt.plot(df['Date'], df['Open'])
plt.xlabel('Date')
plt.ylabel('Open price')
plt.title('Open Price wrt Date')
plt.show()

'Histogram plot'
prices = ['Open','Close', 'High', 'Low']
 
plt.subplots(figsize=(20,10))
 
for i, col in enumerate(prices):
  plt.subplot(2,2,i+1)
  sns.histplot(df[col],kde=True)
plt.show()

C:\Users\IN284372\AppData\Local\Temp\ipykernel_8428\350191310.py:7: MatplotlibDeprecationWarning: Auto-removal of overlapping axes is deprecated since 3.6 and will be removed two minor releases later; explicitly call ax.remove() as needed.
  plt.subplot(2,2,i+1)

'Calculate the 20-days moving average'
df['MA20'] = df['Close'].rolling(window=20).mean() #df.rolling window is used to take K set of data like 1-20,2-21,3-23,....

'Create a line plot of the closing prices and the moving average'
plt.plot(df['Date'], df['Close'], label='Closing price')
plt.plot(df['Date'], df['MA20'], label='20-day MA')
plt.xlabel('Date')
plt.ylabel('Price')
plt.title('Closing price wrt to date')
plt.legend()
plt.show()

'Trend identification using 50-200 Moving Average'

df['MA50']=df['Close'].rolling(window=50).mean()
df['MA200']=df['Close'].rolling(window=200).mean()


plt.figure(figsize=(10,6))
plt.plot(df['Date'], df['Close'], label='Close price' )
plt.plot(df['Date'],df['MA50'], label='50-Day MA')
plt.plot(df['Date'],df['MA200'], label='200-Day MA')
plt.xlabel('Date')
plt.ylabel('Price')
plt.title('Close Price wrt Date')
plt.legend()

'Identify bullish and bearish trends'
bullish = df['MA50'] > df['MA200']
bearish = df['MA50'] < df['MA200']
plt.fill_between(df['Date'], df['Close'], where=bullish,  color='green', alpha=0.3) #filling green color where market is bullish
plt.fill_between(df['Date'], df['Close'], where=bearish,  color='red', alpha=0.3) # filling red color where market is bearish                 

<matplotlib.collections.PolyCollection at 0x279c02b1910>

'Drawing Candle stick chart using  plotly'
import plotly.graph_objects as go
import plotly.io as pio # to save graph

fig = go.Figure(data=[go.Candlestick(x=df['Date'],open=df.Open, high=df['High'],low=df['Low'], close=df['Close'])])

fig.update_layout(xaxis_rangeslider_visible=True,title='NIFTY-50',
    yaxis_title='Prices',xaxis_title='Date')
    
fig.show()
pio.write_image(fig, file='bar.png')

'Creating Box plot to identify outliers'
fig, axs = plt.subplots(nrows=4, ncols=2, figsize=(14, 10))
# Creating a boxplot for each column on a different subplot
sns.boxplot(y=df['Open'],ax=axs[0,0])
sns.boxplot(y=df['Low'], ax=axs[0, 1])
sns.boxplot(y=df['High'], ax=axs[1, 0])
sns.boxplot(y=df['Close'], ax=axs[1, 1])
sns.boxplot(y=df['Range'], ax=axs[2, 0])
sns.boxplot(y=df['MA20'], ax=axs[2, 1])
sns.boxplot(y=df['MA50'], ax=axs[3, 0])
sns.boxplot(y=df['MA200'], ax=axs[3, 1])
fig.suptitle('Boxplots')
plt.show()

From the above boxplot we can see that Range column has more outliers

'Removing outlier from Range column using IQR method'

q1, q3 = np.percentile(df['Range'], [25, 75])
iqr = q3 - q1
lower_bound = q1 - (1.5 * iqr)
upper_bound = q3 + (1.5 * iqr)
f_range = df[(df['Range'] > lower_bound) & (df['Range'] < upper_bound)]
df['Range']=f_range['Range']

'Replacing null values with median function'
mean=df['Range'].mean()
df['Range']=df['Range'].fillna(mean)
df.Range.isnull().sum()

0

'remove one outlier'
df['Range']=df['Range'].replace(342.75,mean)

'Check for outliers in range column'
sns.boxplot(df['Range'])
plt.show()

'Co-relation between each atributes of dataset'
plt.figure(figsize=(10,7))
sns.heatmap(df.corr(),annot=True)
plt.title('Heatmap corelation')
plt.show()

C:\Users\IN284372\AppData\Local\Temp\ipykernel_8428\1575853497.py:3: FutureWarning:

The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.

from above heatmap: Open,Close,High,Low,MA20,MA50 is strongly positive corelated to each other. Range is negatively corelated to all atributes MA200 is strongly negative corelated to 50MA and MA20

Corelation graphs

'Creating a scatterplot for each column on a different subplot'
fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(14, 10))

sns.scatterplot(x=df['Open'],y=df['High'],ax=axs[0,0])
sns.scatterplot(x=df['Open'],y=df['Low'], ax=axs[0, 1])
sns.scatterplot(x=df['Open'],y=df['Close'], ax=axs[1, 0])
sns.scatterplot(x=df['Open'],y=df['MA20'], ax=axs[1, 1])
sns.scatterplot(x=df['Open'],y=df['MA50'], ax=axs[2, 0])

fig.suptitle('Scatter plot')
plt.show()

Machine Learning

'Define feature and target variable'
X = df[['Open', 'High', 'Low', 'Close', 'Range']]
y = df['Close']

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

'transform data'
from sklearn.preprocessing import MinMaxScaler

# Create a scaler object
scaler = MinMaxScaler()

# Fit the scaler on the training data
scaler.fit(X_train)

# Transform the training and testing data
X_train_scaled = scaler.transform(X_train)
X_test_scaled = scaler.transform(X_test)

'Linear Rigression'
from sklearn.linear_model import LinearRegression
# train the linear regression model
model = LinearRegression()
model.fit(X_train, y_train)
y_pred=model.predict(X_test)

# print the model coefficients
print(model.coef_)

[-2.14692184e-16 -1.33226763e-15  1.49880108e-15  1.00000000e+00
  2.09597964e-15]

'Model evaluation'
from sklearn.metrics import mean_absolute_error, mean_squared_error,accuracy_score
print('Mean Absolute Error:', mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(mean_squared_error(y_test, y_pred)))

Mean Absolute Error: 2.158123021156101e-13
Mean Squared Error: 3.925602907031318e-25
Root Mean Squared Error: 6.265463196788661e-13